Cloud Events Exporter¶

The Cloud Events Exporter monitors Kubernetes events and exposes them as Prometheus metrics, enabling monitoring and alerting on cluster events such as pod failures, scaling events, and other operational activities.

Overview¶

Cloud Events Exporter collects Kubernetes events from specified namespaces and converts them into metrics that can be scraped by Prometheus. This enables monitoring teams to:

Track event patterns and trends
Alert on critical events (warnings, errors)
Monitor cluster health through event analysis
Create dashboards showing event statistics

The exporter supports filtering by event type, kind, namespace, and other attributes to reduce noise and focus on relevant events.

Configuration Parameters¶

Field	Description	Scheme
install	Allows to enable or disable deployment of cloud-events-exporter.	bool
name	A deployment name for cloud-events-exporter. Used in deployment name and labels.	string
image	Docker image for cloud-events-exporter. If not specified, uses default image.	string
port	Port to scrape cloud events metrics.	int
namespaces	List of namespaces to monitor for events. Empty list monitors all namespaces.	list[string]
rbac.createClusterRole	Allow creating ClusterRole. If set to `false`, ClusterRole must be created manually.	bool
rbac.createClusterRoleBinding	Allow creating ClusterRoleBinding. If set to `false`, ClusterRoleBinding must be created manually.	bool
filtering	Event filtering configuration to include/exclude specific events.	object
filtering.match	List of event patterns to include. Events matching these patterns will be exported as metrics.	list[object]
filtering.match[N].type	Event type to match (e.g., "Warning", "Normal").	string
filtering.match[N].kind	Resource kind to match (e.g., "Pod", "Deployment").	string
filtering.match[N].reason	Event reason to match.	string
filtering.match[N].message	Event message pattern to match (supports regex).	string
filtering.match[N].namespace	Namespace to match events from.	string
filtering.match[N].reportingController	Reporting controller to match.	string
filtering.match[N].reportingInstance	Reporting instance to match.	string
filtering.exclude	List of event patterns to exclude. Events matching these patterns will NOT be exported as metrics.	list[object]
filtering.exclude[N].type	Event type to exclude (e.g., "Normal").	string
filtering.exclude[N].reason	Event reason to exclude (supports regex, e.g., "Completed OR Pulled OR Started").	string
filtering.exclude[N].message	Event message pattern to exclude (supports regex).	string
serviceMonitor	Service monitor configuration for scraping metrics.	object
serviceMonitor.install	Enables ServiceMonitor creation for automatic metrics scraping.	bool
serviceMonitor.interval	Scraping interval for metrics collection.	string
serviceMonitor.scrapeTimeout	Timeout for scraping metrics. Must be less than interval.	string
serviceMonitor.metricRelabelings	Metric relabeling rules applied to scraped metrics.	list[object]
serviceMonitor.relabelings	Target relabeling rules applied before scraping.	list[object]
extraArgs	Additional arguments for cloud-events-exporter container.	list[string]
extraEnvs	Additional environment variables for cloud-events-exporter container.	object
resources	Resource requests and limits for the container.	v1.ResourceRequirements
securityContext	SecurityContext holds pod-level security attributes.	*v1.PodSecurityContext
nodeSelector	Defines which nodes the pods are scheduled on.	map[string]string
tolerations	Tolerations allow the pods to schedule onto nodes with matching taints.	[]v1.Toleration
affinity	Pod scheduling constraints.	*v1.Affinity
annotations	Annotations to be added to pods.	map[string]string
labels	Labels to be added to pods.	map[string]string
priorityClassName	PriorityClassName assigned to the pods to prevent them from evicting.	string
serviceAccount.install	Specifies whether a ServiceAccount should be created.	bool
serviceAccount.annotations	Annotations for the ServiceAccount.	map[string]string
serviceAccount.labels	Labels for the ServiceAccount.	map[string]string

Metrics¶

The cloud-events-exporter exposes the following metrics:

cloud_events_total{type, kind, reason, namespace, ...} - Total count of events by type, kind, reason, and namespace
cloud_events_warning_total{...} - Total count of warning events
cloud_events_normal_total{...} - Total count of normal events
cloud_events_reporting_controller_warning_total{controller, ...} - Warning events by reporting controller
cloud_events_reporting_controller_normal_total{controller, ...} - Normal events by reporting controller

Configuration Examples¶

Basic Configuration¶

Enable cloud-events-exporter with default settings:

cloudEventsExporter:
  install: true
  name: cloud-events-exporter
  port: 9999
  serviceMonitor:
    install: true
    interval: 30s
    scrapeTimeout: 10s

Monitor Specific Namespaces¶

Monitor events only from specified namespaces:

cloudEventsExporter:
  install: true
  namespaces:
    - monitoring
    - kube-system
    - default

Event Filtering¶

Filter events to focus on warnings and exclude noisy normal events:

cloudEventsExporter:
  install: true
  filtering:
    match:
      - type: "Warning"
        kind: "Pod|Deployment"
      - type: "Warning"
        namespace: "production"
    exclude:
      - type: "Normal"
        message: ".*image.*"
      - reason: "Completed|Pulled|Started"

Custom Resource Configuration¶

Configure resource limits and scheduling preferences:

cloudEventsExporter:
  install: true
  resources:
    limits:
      cpu: 100m
      memory: 128Mi
    requests:
      cpu: 50m
      memory: 64Mi
  nodeSelector:
    node-role.kubernetes.io/worker: "true"
  tolerations:
    - key: "monitoring"
      operator: "Equal"
      value: "true"
      effect: "NoSchedule"

Grafana Dashboard¶

The cloud-events-exporter includes a pre-built Grafana dashboard that provides:

Event trends by type (Warning/Normal)
Top events by namespace and object
Event breakdown by reporting controllers
Cluster vs namespace-scoped events
Event frequency over time

The dashboard is automatically deployed when install: true is set.

Troubleshooting¶

Common Issues¶

No metrics appearing
Check that install: true is set
Verify RBAC permissions for events access
Check pod logs for connection errors
Too many metrics
Configure filtering.exclude to reduce noise
Limit namespaces to specific ones
Use metric relabeling in ServiceMonitor
Missing events
Check filtering.match configuration
Verify namespace access permissions
Review event retention in cluster

Debug Commands¶

# Check cloud-events-exporter pod status
kubectl get pods -l app.kubernetes.io/component=cloud-events-exporter -n monitoring

# View exporter logs
kubectl logs -l app.kubernetes.io/component=cloud-events-exporter -n monitoring

# Check metrics endpoint
kubectl port-forward svc/cloud-events-exporter 9999:9999 -n monitoring
curl http://localhost:9999/metrics

Security Considerations¶

The exporter requires cluster-wide get, list, and watch permissions on events
Consider limiting namespace access by configuring namespaces parameter
Use security contexts to run with minimal privileges
Filter sensitive events that might contain confidential information